KR20060008254A - Infrared sensing system of fire and its sensing method reflecting dynamic pattern of flame - Google Patents

Abstract

The present invention relates to an infrared fire detection method and system using the dynamic characteristics of the flame that can minimize the possibility of error and improve the early detection and accuracy of the fire by processing and analyzing the infrared image using the dynamic characteristics of the flame.

Conventionally, it is not useful for early detection of fire because it detects only a certain range of areas where the temperature is high among infrared images, and it is not useful for early detection of fire, and reflection of sunlight, high temperature artifacts, lights such as cars or street lamps, etc. The reliability of the result is lowered. As a result, a separate visual verification process can not only entail a problem that the efficiency of the system is lowered and a loophole is caused in the rapid response system.

Accordingly, the present invention quantifies the dynamic characteristics of flames such as flame movements and diffusion phenomena and reflects them in the infrared image discrimination method of the fire monitoring system, thereby significantly reducing the false detection rate and automatically detecting and alarming the fire early. To provide a detection method and system, it is necessary to provide a threshold value that quantifies the dynamic characteristics of a flame, a granular determination of blocks, and a fire hazard zone that reflects a flame spread model. to provide.

Infrared image, fire detection

Description

Infrared Sensing System of Fire and its Sensing Method reflecting Dynamic Pattern of Flame}

1 is a block diagram of an infrared fire detection system using the dynamic characteristics of the flame according to the present invention.

2 is a flow chart of the infrared fire detection method using the dynamic characteristics of the flame according to the present invention.

3 is an exemplary view of setting a fire candidate area according to the present invention.

<Description of the major reference numerals>

110: image buffering processor 120: filtering processor

130: block average brightness processing unit 140: fire candidate block processing unit

150: block name change processing unit 160: fire discrimination unit

The present invention relates to a fire detection system using an infrared image. In detail, the dynamic characteristics of the flame can be minimized by improving the accuracy of the error and improving the early detection and accuracy of the fire by processing and analyzing the infrared image using the dynamic characteristics of the flame. An infrared fire detection method and system using the characteristics.

Infrared cameras are widely used in fire surveillance systems because of their ability to provide relatively accurate data at night.However, the same infrared image can be differentiated according to the method of reading the captured infrared image. The fire monitoring system according to the present invention can be said that the method of reading an infrared image determines the performance of the system.

Conventional methods of reading infrared images include: 1) a method of monitoring the fire by a user directly monitoring the naked eye, and 2) comparing an infrared image with a reference image to determine a fire when an abnormal region other than a fixed region identified in advance is found. Or 3) comparing an infrared image with a reference image to find a rapidly changed region, and determining a fire if there is a movement of the changed region.

The method of 1) is a method that the monitoring cost increases and the monitoring effect is determined according to the user's will because the user has to observe 24 hours with the naked eye. The method of 2) and 3) is not useful for initial detection of fire because it finds only a range of areas where the temperature is high in the infrared image and judges it as a fire. The same light, concrete or metal structures, roads, large rocks, etc. can act as error elements and be displayed on infrared images as objects with high temperature, resulting in a very high false detection rate, resulting in less reliable results. Therefore, the user's naked eye must be accompanied by a check, which leads to a decrease in the efficiency of system operation and a loophole in a quick response system.

The present invention has been made to solve the above-described problems of the prior art that does not take into account the characteristics of the flame, the flame generated when the fire occurs, unlike the error element, as the fire or forest fire spreads or the flame moves by the wind The dynamic range of the change in brightness value of the fire zone between the input infrared image frames is relatively large. Using this, it is possible to quickly and accurately fire regardless of weather or climatic environment such as day, night, cloudy day or sunny day. It can be detected.

Therefore, the present invention quantifies the dynamic characteristics of flames such as flame movements and diffusion phenomena and reflects them in the fire monitoring system, thereby significantly reducing false detection rate and automatically detecting and alarming fires early and using infrared flames using dynamic characteristics of flames. An object of the present invention is to provide a detection method and system.

In order to achieve the above technical problem, the present invention provides a fire monitoring system using an infrared image, the image buffering processing unit for updating and storing the infrared image in the image buffer at regular intervals, including the image buffer; A reference image block average brightness processor for selecting a reference image based on a predetermined reference from the infrared images and subdividing the reference image into a plurality of blocks to calculate block average brightness for each block; A reference image fire candidate block processor for determining a fire candidate block by comparing the block average brightness value and a threshold value and specifying a fire candidate area based on the fire candidate block; A block brightness change processor for calculating a block mean of difference (BMD) value by averaging the brightness change values of the fire candidate region for subsequent images including the reference image; It provides an infrared fire detection system using the dynamic characteristics of the flame, characterized in that it comprises a fire determination unit for determining and alarming the fire zone by comparing the average brightness fluctuation value and the threshold value.

The present invention provides a fire detection method using an infrared image, the method comprising: a) an image buffering step of acquiring and storing images at regular time intervals; b) determining a reference image among the infrared images and subdividing the reference image into a plurality of blocks; c) obtaining a Block Mean of Luminance (BML) value for each block; d) determining the fire candidate block by comparing the block average brightness value with a corresponding threshold value; e) specifying a fire candidate area based on the fire candidate block; f) calculating a Block Mean of Difference (BMD) value by averaging the brightness fluctuation values of the fire candidate region for subsequent images including the reference image; g) comparing the block brightness fluctuation value with a corresponding threshold value to determine whether or not a fire zone exists for a fire candidate area, and generating an alarm when the fire zone is determined. Repeatedly according to, steps b) to e) are performed only for the reference image selected at a specific period, and steps c) to e) are performed for each subdivided block of the reference image, and steps f) and g) are Provided is an infrared fire detection method using the dynamic characteristics of the flame, characterized in that performed in the case of determining the fire candidate block in the step).

In addition, the infrared fire detection method using the dynamic characteristics of the flame is implemented on the recording medium in a series of instructions to provide a computer-readable recording medium that can be read and executed by the fire detection system using the infrared image.

Hereinafter, with reference to the accompanying drawings showing an embodiment will be described in detail the technical spirit of the present invention.

1 is a block diagram of an infrared fire detection system using the dynamic characteristics of the flame according to the present invention, the infrared image acquisition apparatus 100, the image buffering processor 110, the filtering processor 120, the block average brightness processor 130, a fire candidate block processor 140, a block brightness change processor 150, a fire determiner 160, and an image buffer 170.

The infrared image obtaining apparatus 100 is an infrared camera that captures an infrared image and transmits the infrared image to the image buffering processor 110. The infrared image obtaining apparatus 100 is fixed to the monitoring target region or configured to be rotated or vertically adjusted by a control system.

The image buffering processor 110 includes an image buffer 170, and stores the image data in the image buffer 170 at predetermined time intervals from the image string acquired by the infrared camera.

In this case, the noise removing operation may be necessary to improve the discrimination of the acquired infrared image. The filtering processor 120 reads image data stored in the image buffer 170 to median a filter. Remove the impulsive noise included in the image using.

The block average brightness processor 130 selects a reference image at regular intervals from among the images, divides the selected reference image into a plurality of macroblocks, and calculates an average brightness value for each block using brightness values of pixels constituting the block.

The fire candidate block processor 140 determines a block having a value greater than or equal to the fire candidate block threshold by comparing the block average brightness value and the fire candidate block threshold value set based on the dynamic characteristics of the flame. Determine the fire candidate area extended to the dynamic characteristics of the flame around the block.

The block brightness variation processor 150 calculates and averages the block brightness variation by repeatedly calculating and averaging the average brightness variation of the pixels corresponding to the fire candidate region for successive images after the corresponding reference image. To provide contrast data that can be compared to characteristics.

The fire determination unit 160 compares the block brightness change value with a corresponding threshold value and determines a fire zone and outputs an alarm if the threshold value is equal to or greater than the threshold value.

2 is a flowchart illustrating an infrared fire detection method using a dynamic characteristic of a fire according to the present invention. The image buffering step F200, the filtering step F210, the reference image setting and blocking step F220, and the block average brightness are shown. (BML) calculating step (F230), fire candidate block determination step (F240), fire candidate area setting step (F250), block name change value (BMD) calculation step (F260), and fire determination and alarm step (F270) Include.

In describing each of the above steps, if necessary, it will be described in conjunction with each component shown in FIG.

The image buffering step F200 stores the infrared image signal photographed by the infrared image acquisition apparatus 100 continuously (e.g. 1 second period) in the image buffer 170. At this time, the analog video signal input by the infrared camera is converted into a digital video signal by an image grabber (not shown) and output.

If it is necessary to remove the noise of the infrared image obtained in the above step, the filtering step (F210) is added to perform the filtering to remove the impulsive image noise on the image signal stored in the image buffer 170. In this case, various types of noise removing filters may be used as the filter, but a median filter is generally used.

In the reference image setting and blocking step (F220), a reference image for calculating a block average brightness is set and blocked, wherein the reference image is an image that is an object for determining whether an actual fire occurs among infrared images, and the reference image is used. The reason for the selection is to perform fire discrimination on only a part of the acquired infrared images to reduce the load on the system and maximize the efficiency of the system.

The selection criteria of the reference image may be any, but a certain time period may be applied in consideration of the ease and efficiency of implementation. That is, in the state where the reference image is not initially set, the buffered first image may be set as the reference image, and thereafter, the reference image may be updated and set at a predetermined period (eg 30 seconds). Although it may coincide with the infrared image acquisition period of F200, in consideration of the load of the system, it is preferable to select the reference image at a period larger than the infrared image acquisition period as described above.

Blocking means subdividing the reference image into a plurality of blocks (macro blocks) of a constant size (e.g. 8x8 or 16x16 pixels), which enables early detection of small flames.

In the calculation of the block average brightness (BML) of the reference image (F230), a block average brightness (BML) of each block for the set reference image is obtained, which is calculated by Equation 1 by the block average brightness processor 130.

(수식 1)

(Formula 1)

Here, M1 represents a row size of a block, N1 represents a column size of a block, and P (m, n) represents a brightness value for each pixel of the block.

In the fire candidate block determination step (F240) of the reference image, when the block average brightness value (BML) of the step (F230) is compared with the threshold value reflecting the brightness characteristics of the flame, if a block average brightness value that is equal to or greater than the threshold value is found, the fire occurs. Select as a candidate block. In this case, if it is less than the threshold value, the process branches to the reference image setting and blocking step F220 to obtain a new reference image.

In the step of setting the fire candidate area of the reference image (F250), the area where the flame can be spread, that is, the fire candidate area, is specified around the fire candidate block, and the flame is mainly moved upward and diffused from side to side to reflect the dynamic characteristics. The area including the blocks located on the left, right, and top of the fire candidate block may be selected as the fire candidate area. Figure 3 shows an embodiment of the selection of the fire candidate area, the block is selected in a 3 x 3 arrangement structure, the fire candidate block is located in three rows 2 columns (block 8), the flame is spread around the fire candidate block It reflects the dynamic characteristics of

In the block brightness variation value (BMD) calculation step (F260) for the fire candidate region, the fire candidate regions are mapped to a plurality of consecutive images (eg 30) after the reference image including the reference image, and the fire candidate regions for each image are displayed. The block brightness fluctuation value BMD, which is an average value of the brightness changes, is calculated, and the value is calculated by Equation 2 by the block brightness fluctuation processing unit 150.

(수식 2)

(Formula 2)

Where M2 is the row size of the fire candidate area, N2 is the column size of the block, Pt- 1 (m, n) is the pixel brightness value of the previous frame, and Pt (m, n) is the pixel brightness value of the current frame.

In the fire discrimination and warning step F270, the block brightness change value calculated in the step F260 is compared with a corresponding threshold value to determine a fire zone and generate an alarm when the threshold value is greater than or equal to the threshold value. This is to provide a threshold value in consideration of the dynamic characteristics in the spread of the flame, and determine whether the measured variation value (BMD) is equal to or greater than the threshold value to determine whether the flame characteristic is met. In this case, if the threshold value is smaller than the threshold value, the process branches to the reference image setting and blocking step F220 to obtain a new reference image.

The present invention can also be embodied as computer readable code on a computer readable recording medium, which can be read and executed by a fire detection system using an infrared image. Computer-readable recording media include all types of recording systems that store data that can be read by a computer system. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage system, and the like, which are also implemented in the form of carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. And functional programs, codes, and segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

Although specific terms have been used in the foregoing description, they are used only for the purpose of describing the present invention and are not used to limit the scope of the present invention as defined in the meaning or claims. In addition, preferred embodiments of the present invention are disclosed for the purpose of illustration, those skilled in the art will be possible to various modifications, changes, additions, etc. within the spirit and scope of the present invention, if such modifications and changes are obvious to those skilled in the art the present invention It should be considered that it is included in the technical scope of the.

In particular, the infrared image acquisition period, the reference image acquisition period, the size of the block, and the setting range of the fire candidate area are various variations, and it is necessary to select an optimal value in the trade-off between the accuracy of the results and the system load. The example values presented can be referenced.

According to the present invention, by setting the threshold value and the fire hazard area by using the dynamic characteristics of the flame to distinguish the infrared image, it is possible to dramatically improve the accuracy of the fire detection even in the presence of a large number of error elements, block unit The detection system has the advantage of excellent detection accuracy even in the early stage of ignition.

Therefore, since the alarm signal is automatically generated upon detection of fire based on reliable data, it is possible to effectively and quickly monitor the system, thereby preventing the enormous human and physical damage at low cost.

Claims (10)

In the fire detection method using an infrared image, a) an image buffering step of acquiring and storing images at regular time intervals; b) determining a reference image and subdividing the reference image into a plurality of blocks; c) obtaining a Block Mean of Luminance (BML) value for each block; d) determining the fire candidate block by comparing the block average brightness value with a corresponding threshold value; e) specifying an area including a block located on the left, right, and upper sides of the fire candidate block as a fire candidate area; f) calculating a Block Mean of Difference (BMD) value by averaging the brightness fluctuation values of the fire candidate region for subsequent images including the reference image; g) comparing the block brightness fluctuation value with a corresponding threshold value to determine whether a fire zone is a fire zone and generating an alarm when the fire zone is determined, wherein the step a) comprises inputting an infrared image. Repeated according to the cycle, steps b) to e) are performed only for a reference image selected at a predetermined cycle, and steps c) to e) are performed for each subdivided block of the reference image, and steps f) and g). Infrared fire detection method using the dynamic characteristics of the flame, characterized in that is carried out if it is determined in step d) as the fire candidate block. The method of claim 1, wherein the step a) further comprises the step of performing a filtering by means of a median filter to remove impulsive noise on the acquired infrared image. Infrared fire detection method. The method of claim 1, wherein the reference image is selected at a predetermined time period in step b), and the size of the block is 8x8 or 16x16 pixels. The method of claim 1, wherein the block average brightness (BML) of the step c) is based on the following equation: infrared fire detection method using the dynamic characteristics of the flame

Here, M1: row size of the block, N1: column size of the block, P (m, n): brightness value for each pixel of the block. According to claim 1, wherein the fire candidate area in step e) is a block of 3 x 3 array structure, wherein the fire candidate block is located in three rows and two columns, infrared fire detection using the dynamic characteristics of the flame. Way. The method of claim 1, wherein the BMD value of step f) is determined by the following equation.

Where M2: row size of the fire candidate area, N2: column size of the block, Pt- 1 (m, n): pixel brightness value of the previous image, Pt (m, n): pixel brightness value of the current image. A fire monitoring system using an infrared image, the fire monitoring system comprising: an image buffering processor configured to update and store an infrared image at a predetermined cycle, including an image buffer; A block average brightness processor configured to calculate a block average brightness for each block by determining a reference image based on a predetermined reference among the acquired infrared images and subdividing the reference image into a plurality of blocks; A fire candidate block processor for determining a fire candidate block by comparing the block average brightness value and a threshold value and specifying a fire candidate area based on the fire candidate block; A block brightness change processor for calculating a block mean of difference (BMD) value by averaging the brightness change values of the fire candidate region for subsequent images including the reference image; Infrared fire detection system using the dynamic characteristics of the flame, characterized in that it comprises a fire determination unit for comparing the average brightness fluctuation value and the threshold value to determine the fire zone, and generates and outputs an alarm. The infrared fire detection system of claim 7, further comprising a filtering processor configured to remove impulsive noise by using a median filter on the infrared image acquired by the image buffering processor. A recording medium having recorded thereon a program which can be read and executed by an infrared fire monitoring system. a) an image buffering step of acquiring and storing images at regular time intervals; b) determining a reference image and subdividing the reference image into a plurality of blocks; c) obtaining a Block Mean of Luminance (BML) value for each block; d) determining the fire candidate block by comparing the block average brightness value with a corresponding threshold value; e) specifying an area including a block located on the left, right, and upper sides of the fire candidate block as a fire candidate area; f) calculating a Block Mean of Difference (BMD) value by averaging the brightness fluctuation values of the fire candidate region for subsequent images including the reference image; g) comparing the block brightness fluctuation value with a corresponding threshold value to determine whether a fire zone is a fire zone and generating an alarm when the fire zone is determined, wherein the step a) comprises inputting an infrared image. Repeatedly according to the cycle, steps b) to e) are performed only for the reference image selected at a predetermined cycle, and steps c) to e) are performed for each subdivided block of the reference image, f) and g). And the step is performed if it is determined in step d) that it is a fire candidate block. 10. The recording medium of claim 9, wherein the step a) further comprises a filtering step for removing impulsive image noise from the acquired infrared image.

Images